frei0r | A large collection of free and portable video plugins | Video Utils library

The fundamental issue in these audio files appears to be the frequently dropped frames (each containing 960 audio samples). There is an instance of 8117 seconds gap between 2 successive frames in the first file. Because the MKA files were formed without filling these dropped frames, they are effectively variable-sampling-rate streams while labeled as constant-sampling-rate. This discrepancy makes your audios to appear shorter than they were recorded, explaining why your output is often much longer than expected and has been wrecking havoc on your attempt to work on these files.

While atm I do not know if FFmpeg offers a mechanism to fix/estimate the dropped frames in these files, yYou can brute-force/ignore the dropped frames by:

amix

Your file seems to be missing some color information:

Try this (the short clip inserted at 3-second mark):

You are using formData with the HTTP PUT method which corrupts the upload as a binary data stream is expected and not MIME data.

To use formdata:

Upload an object with HTML forms

From your log I can see that your input audio is MP3.

The forfiles command is a nasty beast, because there are several caveats:

• it is slow (particularly because it cannot run internal commands of the hosting command prompt);
• it handles wildcards differently than most other commands, hence /M *.* does not match all files but only such with an extension; to really match all files, use /M * or skip it since it is the default anyway;
• it applies backslash-escaping, which is particularly annoying with paths ending in \, like the root directory of a drive /P "D:\", which causes a syntax error since the closing quotation mark is considered as escaped; to work around that, preferably append a . like /P "D:\.", or remove the quotation marks like /P D:\, though this exposes potential whitespaces or special characters to the parser;
• all of the special @-variables that return the path and/or name of iterated items provide the values in quoted manner, which is particularly frustrating when it comes to concatenation;
• it iterates over both files and directories that match the given mask; to distinguish between them you could use the special @isdir variable, but you will need an if statement for this (like if @isdir==FALSE or if @isdir==TRUE), which is an internal cmd.exe command, requiring its explicit instantiation even when you would not need it else;
• handling of the command behind /C and its arguments is terribly implemented, leading to the problem that directly running external commands (so without cmd /C) may fail, unless you are aware of the mostly working fix by stating the command name twice (like /C "command.exe command.exe --parameter argument");
• even its basically nice /D option (which is the only reason why forfiles might suit better than for) to filter for the relative last modification date (but not time) is badly implemented when a positive number (like /D +1) is used, because this uselessly points to the future;

All of these issues lead me to the point that I suggest not to use forfiles and to use a standard for loop instead, like this (note also the changed mask *.flac):

• In a batch-file:

The source of the differences is that FFprobe counts the discarded packets, and FFmpeg doesn't count the discarded packets as frames.

Your results are consistent with video stream that is created with 3 B-Frames (3 consecutive B-Frames for every P-Frame or I-Frame).

According to Wikipedia:

I‑frames are the least compressible but don't require other video frames to decode.
P‑frames can use data from previous frames to decompress and are more compressible than I‑frames.
B‑frames can use both previous and forward frames for data reference to get the highest amount of data compression.

When splitting a video with P-Frame and B-Frame into segments without re-encoding, the dependency chain breaks.

• There are (almost) always frames that depends upon frames from the previous segment or the next segment.
• The above frames are kept, but the matching packets are marked as "discarded" (marked with AV_PKT_FLAG_DISCARD flag).

For the purpose of working on the same dataset, we my build synthetic video (to be used as input).

Building synthetic video with the following command:

I moved the creation of the MediaRecorder instance to the recurrent function and now it is working properly:

Your input is detected as interlaced but with half-height (288) and double the framerate (50). This may be due to missing or unrecognized boundary markers in the JPEG2000 packets. I assume this is meant to be PAL --> 720x576@25i.

Try using the tinterlace filter first to interleave the input "frames" to double-height and half fps, and then continue the original sequence of filters.

"tinterlace=mode=merge,bwdif=0,select='if(eq(n\,0),1,floor(t)-floor(prev_selected_t))'"